The Commons

Five stories appeared on Techmeme today. None about the same company. None about the same industry. All about the same structural failure: open source repositories, visual platforms, web archives, email protocols, and the culture that sustains them are degrading in the same way, at the same time, for the same reason. AI made contributing to shared spaces free. It did not make maintaining them free. The gap is swallowing the commons.

The Code

VLC and Blender maintainers are reporting a decline in the average quality of contributions, coinciding with the rise of AI coding tools. TechCrunch framed the asymmetry precisely: building new features is easier, but maintaining them is just as hard.

This is the structural problem. AI made writing code cheap. It did not make reviewing code cheap, debugging code cheap, or making the architectural decisions that keep a codebase coherent cheap. The volunteers who maintain open source projects — the people who built the infrastructure most of the tech industry runs on — are now drowning in pull requests from contributors who can generate code but cannot evaluate it.

A Stanford study found in 2022 that programmers using AI coding assistants tended to produce less secure code. The barrier to entry fell. The quality floor fell with it.

The Feed

Pinterest's degradation follows a twelve-month arc of attempting to solve a problem with the technology creating it.

• MAR 2025
  Pinterest announces tools to label AI-generated content, after users begin asking whether the platform is dead.
• May 2025 Pinterest rolls out an "AI modified" label globally.
• Jul 2025 Financial Times reports Pinterest is "increasingly filled with AI interior design images."
• Oct 2025 Pinterest adds tools letting users limit AI content in their feeds.
• JAN 2026
  SEC filing: Pinterest plans to lay off less than 15% of its staff.
• Feb 4 Pinterest fires two employees who wrote custom software to manipulate its AI content detection system.
• FEB 2026
  404 Media: artists say the platform has "gotten worse" despite every intervention.

One user described the current state: ninety-five out of a hundred pins are AI slop. Hand-drawn artwork from a decade ago is being flagged as "AI modified" by Pinterest's own detection tools. The platform's original value — human curation of visual references — has been buried under machine-generated noise.

And the company's response to the flood was to lay off staff and invest more in the technology causing the flood. Reddit moderators are reporting the same pattern: AI slop is eroding community trust, and moderation cannot keep pace with generation.

The Archive

English-language Wikipedia banned Archive.today this week after editors discovered the archiving service had been used to direct a DDoS attack against Wikipedia and had tampered with its own snapshots — altering archived pages to frame a blogger who had investigated the service's operators.

The service that archived the web's pages was corrupting them. Meanwhile, Google announced it is ending Gmailify and POP access in Gmail. POP — the Post Office Protocol — was defined in 1984. For forty-two years it allowed any email client to connect to any email server. Google is closing that door. Open protocols are becoming attack surfaces too expensive to defend.

One protocol corrupted its own archive. Another closes after four decades. Both point the same direction: open infrastructure is becoming harder to sustain than the value it creates.

The Culture

Harper's published a profile of Cluely founder Roy Lee and a generation of "highly agentic" young men for whom agency has replaced technical skill as Silicon Valley's most valued trait. Lee was suspended from Columbia for using AI to cheat on a technical interview. He filmed himself doing it, posted the footage online, and raised tens of millions in venture funding.

I do not obtain value from reading books. The future won't reward effort. It'll reward leverage.

Another founder profiled in the piece took business calls from his high school bathroom by bribing drug dealers for hall passes, then subcontracted coding jobs to overseas freelancers while keeping the markup. The piece's featured subjects treat friction — learning, building, understanding — as the thing to eliminate. This is not a bug report about AI tools. It is an ideology: quality gates are obstacles, not standards.

The Price

Here is the number that ties it together. Source Global reports the US consulting market is set to grow 7% in 2026 — the fastest pace in the post-COVID era. Companies are paying more for human expertise because AI generates so much noise that sorting signal from slop requires people who know the difference.

The commons — open source, open protocols, shared platforms, public archives — was sustained by a cost equilibrium. Contributing was hard enough that only people who could also maintain bothered to contribute. AI broke that equilibrium. Contributing is now free. Maintaining is not.

Friction hasn't disappeared. It's been repriced.

What used to be free — community maintenance, open protocols, human curation — is becoming expensive: consulting fees, enterprise support contracts, premium filtering. The commons isn't being enclosed by gatekeepers. It's being enclosed by the absence of gates.